Gait Biomarkers Classification by Combining Assembled Algorithms and Deep Learning: Results of a Local Study.

Area of Science:

• Artificial Intelligence
• Biomedical Engineering
• Data Science

Background:

• Diabetic neuropathy (DN) affects a significant global population.
• Early recognition of DN is crucial for effective management.
• Machine learning offers advanced pattern recognition for medical diagnoses.

Purpose of the Study:

• To develop and validate a machine learning model for diabetic neuropathy recognition.
• To investigate the efficacy of gait biomarkers in identifying DN.
• To leverage AI for improved healthcare decisions in DN detection.

Main Methods:

• Utilized a homemade body sensor network to collect gait data from individuals with and without DN.
• Processed walking pattern data using three sampling criteria and 23 classifiers.
• Employed a deep learning algorithm, optimizing its architecture for enhanced performance.

Main Results:

• Achieved a classification accuracy exceeding 85% for DN recognition.
• Demonstrated the effectiveness of combined machine learning approaches and gait analysis.
• Validated the model on a specific population sector in southern Mexico.

Conclusions:

• The developed machine learning model shows high accuracy in identifying diabetic neuropathy.
• Gait biomarkers captured via body sensor networks are valuable for DN detection.
• This AI-driven approach supports precise and minimally invasive diagnostic capabilities for DN.